From Design to Construction: Lessons Learned about Bioretention Performance
The Edna Bailey Sussman Foundation Internship Report
December 2015
Tom Baker, MLA Candidate
State University of New York College of Environmental Science and Forestry
Department of Landscape Architecture
Background
Bioretention systems have become one of the most widely used stormwater best
management practices (BMP) since its development in the early 1990’s in Prince George’s
County, Maryland. Their implementation is critical for urban environments as studies have
shown their effectiveness in stormwater and combined sewer overflow management (Patwardhan
et al. 2005; Zhen et al. 2006). As such, they are encouraged by many state and local governments
through incentives and are sometimes required through stormwater management ordinances. As
bioretention systems increase visibility in the urban environment and become permanent urban
infrastructure, proper procedure in maintaining visually appealing systems is necessary.
Research, both of field and laboratory studies, have shown the effectiveness of
bioretention systems on reducing stormwater runoff quantity (Chapman and Horner 2010;
Zimmer et al. 2007) and removal of heavy metals and pollutants (Davis et al. 2001; Davis et al.
2003). Further, research has highlighted the importance of vegetation and appropriate plant
selection as a parameter in pollutant/nutrient removal (Lucas and Greenway 2008) and water
absorbing capacity for runoff retention (Bratieres et al. 2008). Though research has shown the
importance of vegetation on bioretention system performance, little research has been conducted
on overall planting success. There remains limited literature available to designers regarding
specific planting recommendations based on bioretention vegetation survivorship.
However, as bioretention systems become more widely used, poor and variable design,
construction, and maintenance practices remain a concern. Improper soil mixture selection and
placement have resulted in poor infiltration rates (Cosgrove and Bergstrom 2003) and unequal
ponding (Carpenter and Hallam 2010). As maintenance practices evolve, researchers have called
for an emphasis on maintenance recommendations (Davis et al. 2009).
In order for bioretention systems to be successful close coordination from design to
construction to maintenance and monitoring needs to be practiced. Lack of communication
between designers and contractors can lead to specifications that call for materials that are
unavailable (e.g. filter media, plant species) and poor oversight of the installation process.
Further, lack of coordination can lead to improper maintenance protocol and poor success of
vegetation.
To understand the relationship between design, construction, and maintenance of
bioretention systems, this study focused on bioretention cells designed and installed under
Onondaga County’s Save the Rain program in central New York. Save the Rain is a
comprehensive stormwater program implemented to reduce pollution into Onondaga Lake and
its tributaries. The program has implemented 175+ green infrastructure projects to reduce
overflow of the combined sewer system during extreme rain events. Over 30 of the projects
include bioretention as the focus of reducing stormwater runoff (Save the Rain).
Study Objectives
Save the Rain provided an opportunity to study the inconsistencies from design,
construction, and maintenance of bioretention systems and how these variables impact vegetation
success. This study was intended to specifically address the following questions: (1) Are soils
installed accurate to the designer’s original specifications?, (2) Do more detailed planting plans
lead to more successful plantings?, and (3) Does a higher maintenance level positively impact
vegetation performance?
Methods
Site Selection
Ten bioretention cells were surveyed in Syracuse, NY during June and July 2015. Save
the Rain program requirements called for soils at all sites to follow the same specifications, so
contractors were varied between sites to understand if soils were installed differently than
specified. Sites were all constructed three to four years prior to this study and varying sizes and
contexts were included to introduce variables in planting design and site conditions.
Field Data Collection and Analysis
To test the soils, all existing mulch and sediment was removed to gain access to the S1
(plant root growing medium) soil layer. The first six inches of the S1 soil was collected, using a
core soil sampler at each site’s inflow, emergency overflow, and two places near the center. At
five sites there were several inflow points and/or emergency overflow locations, so it was
assumed any single point would be characteristic of the drainage area. Soils were air-dried at
room temperature.
Soil samples were analyzed for organic matter, pH, phosphorus, nitrogen, and texture
(sand, silt, clay) at SUNY-College of Environmental Science and Forestry’s Forest Soils
Research Laboratory. Soil analyses were compared to the specified ranges provided by
Onondaga County’s Save the Rain technical documents.
Vegetation assessments were conducted visually at a single point in time for each species
providing an indicator of overall vigor, branch and foliage density, leaf shape and condition. An
additional metric, survivability, was needed for perennials and grasses to measure the planting
success of groups. Survivability was assessed visually to compare observed establishment to
what was intended from planting plans. Separate assessments were completed for species that
had two or more separate and distinct planting groups at a site. Time of year observed, age of
plantings, and initial planting size/age were taken into account for vegetation assessments.
Original planting specifications were reviewed for base information. Trees and ground cover
were not included in the analysis because of small sample sizes.
Maintenance
Maintenance was assessed using Save the Rain maintenance protocol and comparing the
protocol to observable practices at each site. Discussing maintenance practices with Save the
Rain personnel was an objective to understand specific maintenance concerns/successes at
specific sites. Attempts to contact administration personnel responsible for overseeing
maintenance was unsuccessful. Administration was willing to send maintenance technical
documents but were unwilling, or unable, to discuss specific maintenance practices.
Results
Site Parameters
Sites were constructed in 2011 (n=3) and 2012 (n=7), varied in area from 63 m2 to 858
m2, owned by City of Syracuse, and included five different contractors. Contextual land use
included paved (n=6), grass (n=3), and paved/grass (n=1). Seven sites included detailed planting
plans (i.e. single plant species identified to specific planting location) and three sites had nondetailed planting plans (i.e. two or more species assigned to a specific planting location via
hatching) (Table 1). All sites had at least one runoff entry point (inflow) and one emergency
overflow.
Table 1. Study sites and parameters including year completed, area, land use type, planting plan
specifications, and contractor.
Site
Year Completed
Barker
Genesee
Hazard 1
2012
Estimated 2011
2012
Area (sq.
meters)
138
49
72
Hazard 2
Leavenworth 1
Leavenworth 2
Pass 1
Pass 2
South State
Sunnycrest
2012
2012
2012
2011
2011
2012
2012
35
63
59
226
858
43
533
Grass
Paved
Paved
Planting
Plans
Non-detailed
Detailed
Detailed
John R. Dudley Construction
Barrett Paving Materials
Davis Wallbridge
Paved
Paved
Paved
Grass
Grass
Paved
Paved/Grass
Detailed
Detailed
Detailed
Detailed
Detailed
Non-detailed
Non-detailed
Davis Wallbridge
John R. Dudley Construction
John R. Dudley Construction
Economy Paving
Economy Paving
Davis Wallbridge
ACTS II Construction
Land Use
Contractor
Soil Analysis
Organic matter, pH, phosphorus, and texture were found to range from 1-3%, 7.9-8.7, 035ppm, and 83-88% (sand), 7-10% (silt), and 5-7% (clay), respectively; compared to specified
ranges of 2-4%, 5.5-6.5, 20-100ppm, and 78-89% (sand), 6-12% (silt), and 3-6% (clay),
respectively (Table 2).
Table 2. Study sites, original soil specifications, and analyzed soil ranges.
Organic
Matter
(%)
pH
Phosphorus
(ppm)
Sand
(%)
Silt
(%)
Clay
(%)
Specified Range
2-4
5.5-6.5
20-100
83-88
6-12
3-6
Barker
Genesee
2
3
8.1
8.0
0.00
34.51
88
85
7
9
5
6
Hazard 1
Hazard 2
Leavenworth 1
Leavenworth 2
Pass 1
Pass 2
2
2
2
2
1
2
8.1
8.1
7.9
8.1
8.1
8.1
0.00
0.00
0.01
0.00
6.29
11.88
86
86
86
86
86
87
8
8
9
8
8
7
6
6
5
6
7
7
South State
Sunnycrest
2
2
8.7
7.9
0.19
14.09
84
88
10
7
6
5
Vegetation Assessments
At the ten sites, a total of 49 nursery stock species were originally planted including six
species of trees, 10 species of small trees/shrubs, 17 species of perennials, 12 species of grasses,
and four species of groundcover. Vegetation vigor and survivability varied between sites (Figure
1, Figure 2, Figure 3). Of the originally planted nursery stock, six were not observed at the time
of the study, due to either lack of surviving or not being planted. Total species per site ranged
from two to 15. Two species were common among six sites, one common among five sites, four
common among four sites, four common among three sites, 12 common among ten sites, and 26
found only at one site.
Figure 1. Average vigor and survivorship of grasses among sites.
Figure 2. Average vigor and survivorship of perennials among sites.
Figure 3. Average vigor small trees/shrubs among sites.
Discussion
Soils
Installed soils were generally similar to original specifications. Clay content and organic
matter differed from specified ranges by no more than 1%. Measured soil phosphorus ranged
from 0.00 – 34.51 ppm, well below what was originally specified. It is assumed original
phosphorus levels were higher. Soil pH also varied from original specifications, measuring from
7.9 to 8.7. There was a discrepancy between supplied soil and specific soil. Reviewing soil sign
off sheets revealed that the soil supplied had considerably higher pH than specified.
The planting design and species selection seems appropriate for the original soil
specifications, though much variation in plant vigor and survivability was observed among sites.
Variation in plant performance can be attributed, in part, to the original high soil pH as well as
local conditions that exacerbate salt and pH issues.
Vegetation and maintenance/construction
Vegetation survivability and vigor varied, with all sites exhibiting some poor
performance, and with low survivability and vigor commonly associated with edge plantings at
all sites, except Sunnycrest. Edge plantings, regardless of species, showed diminished growth,
low branch/foliage density, and low group survivability when compared to inner region
plantings. Poor performance of edge plantings may be attributed to higher stress environments
related to heat, compaction, and salt, specifically when planted adjacent to hardscapes.
Similar to Cosgrove and Bergstrom (2003) and Chavez et al. (2015), vegetation
survivability varied in relation to maintenance and design specifications. Among sites with poor
maintenance and inadequate site design, species survivability and distribution was low when
compared to original planting specifications. East Genesee and South State showed a lack of post
installation maintenance, specifically in regards to removal of sediment build up. Large amounts
of sediment buildup were evident across the surface East Genesee and South State and around
emergency overflows (Figure 4, Figure 5). Three of seven nursery stock species were observed at
East Genesee including two woody plants (V. angustifolium ‘Top Hat’ and C. caroliniana), with
Figure 4. Lack of sediment removal from South State resulting in homogenous topography and sediment
build-up around site emergency overflow.
Figure 5. Sediment build-up around the inlet of East Genesee, resulting in poor vegetation establishment and
a compacted surface layer.
V. angustifolium ‘Top Hat’ standing but dead, and one grass (S. scoparium). C. caroliniana and
S. scoparium were in very poor health and only one individual of S. scoparium was observed.
The rest of East Genesee was established with volunteer species. Chavez et al. (2015) showed
similar results when reviewing vegetation survivorshop within seven bioretention cells in
Oklahoma. Their 2015 study showed an overall decline in nursery stock species which gave way
to volunteer species recruitment. South State showed consistently low plant vigor and
survivability, which can in part be attributed to sediment build up. South State shows evidence of
water transporting mulch and sediment throughout the site, specifically built up around the
emergency overflow, which created a topographically homogenous site.
Pass 1 and Pass 2 showed low vegetation survivability around inlets and areas of high
erosion. Both sites showed evidence of high water velocity entering the site through inlets and
creating surface preferential flow paths and soil and mulch erosion around inlets (Figure 6,
Figure 7). Pass 2 also exhibited erosion on banks (Figure 8). The cause of erosion was
identifiable to water velocity near inlets, but the cause of erosion on the slopes was unclear. Due
to a serious lack of vegetation and severe displacement of soil, erosion on the slopes was likely
occurring well before the start of this study.
Figure 6. Erosion and preferential surface flow path from inlet at Pass 1, resulting in poor vegetation vigor
and survivability.
Figure 7. Erosion surrounding inlet from Pass 2, resulting in poor vegetation vigor and survivability.
Figure 8. Erosion on bank of Pass 2, resulting in poor vegetation vigor and survivability.
Vegetation survivability and vigor was considerably higher at Sunnycrest than any other
site. Most notably, I. verticillata, which was planted on a hillex in the center of the Sunnycrest
cell, had excellent overall vigor while I. verticillata had poor vigor where found at four other
sites. Vegetation success can in part be attributed to proper planting locations based on
individual species needs.
Maintenance practices were assessed with limited available resources. Overall,
maintenance was poor, specifically in regards to sediment removal, erosion management, and
replacement of large swaths of dead vegetation.
Conclusion
Variable design, construction, and maintenance practices related to overall vegetative
success. Specifically, poor practices were observed in regards to inlet design, vegetation
placement, and removal of sediment. In the case of this study, initial vegetation survivability was
poorly considered with installed soil pH considerably higher than species needs. The findings of
this study call for increased coordination through the phases of design, construction, and
maintenance. Further, concern of low maintenance resources calls for more consideration of less
complex planting designs (fewer number of plant species, ease of maintenance).
Overall, this study shows that lack of coordination between design, construction, and
maintenance exists and negatively impacts the success of vegetation. There is a need for
increased literature for designers regarding proper planting specifications in bioretention
systems.
Further Studies
Further studies will be required to better establish bioretention design, construction, and
maintenance recommendations. Specific studies should address dynamics of filter media
chemistry/composition over time and its impact on vegetation performance, varying maintenance
practices impacts on vegetation performance, and comparisons of vigor and survivorshop
between types of plantings (e.g. grasses vs. perennials).
Acknowledgements
Many thanks go to the Edna Bailey Sussman Foundation as this research would not have
been possible without them. I would also like to thank my major professor, Tim Toland, for his
continued guidance and support through this project. I also thank Mark Burger from Onondaga
County Soil and Water Conservation District for his sponsorship and support through the project.
References
Bratieres, Katia, Tim Fletcher, Ana Deletic, and Yaron Zinger. 2008. Nutrient and Sediment
Removal by Stormwater Biofilters: A Large-scale Design Optimisation Study. Water
Research. 42: 3930-3940.
Carpenter, D.D., and Laura Hallam. 2010. Influence of Planting Soil Mix Characteristics on
Bioretention Cell Design and Performance. Journal of Hydrologic Engineering. 15:40416.
Chapman, Cameron, and Richard R. Horner. 2010. Performance Assessment of a StreetDrainage Bioretention System. Water Environment Research. 82:109-119.
Chavez, Rebecca A., Glenn O. Brown, Reid R. Coffman, and Dan E. Storm. 2015. Design,
Construction and Lessons Learned From Oklahoma Bioretention Cell Demonstration
Project. Applied Engineering in Agriculture. 31:63-71.
Cosgrove, James F. Jr., and Jeremiah D. Bergstrom. 2003. Design and Construction of
Biofiltration Basins: Lessons Learned. World Water & Environmental Resources
Congress 2003. n. pag.
Davis, Allen P., Mohammad Shokouhian, Himanshu Sharma, and Christi Minami. 2001.
Laboratory Study of Biological Retention for Urban Stormwater Management. Water
Environment Research. 73:5-14.
Davis, Allen P., Mohammad Shokouhian, Himanshu Sharma, Christi Minami, and Derek
Winogradoff. 2003. Water Quality Improvement through Bioretention: Lead, Copper,
and Zinc Removal.” Water Environment Research. 75:73-82.
Davis, Allen P., William F Hunt, Robert G. Traver, and Michael Clar. 2009. Bioretention
Technology: Overview of Current Practice and Future Needs. Journal of Environmental
Engineering. 135:109-117.
Lucas, William C., and Margaret Greenway. 2008. Nutrient Retention in Vegetated and
Nonvegetated Bioretention Mesocosms. Journal of Irrigation and Drainage Engineering.
134:613-623.
Patwardhan, Avinash S., J Timothy Hare, Thomas Jobes, and Daniel Medina. 2005. Analyzing
Potential Benefits of Low Impact Development in Reducing Combined Sewers
Overflows. Impacts of Global Climate Change. n. pag.
Save the Rain. Save the Rain: About. (Accessed: May 11, 2015).
Zhen, Jenny, Leslie Shoemaker, John Riverson, Khalid Alvi, and Mow-Soung Cheng. 2006.
BMP Analysis System for Watershed-Based Stormwater Management. Journal of
Environmental Science and Health, Part A. 41:1391-1403.
Zimmer, C.A., I.W. Heathcote, H.R. Whiteley, and H. Schroter. 2007. Low-Impact-Development
Practices for Stormwater: Implications for Urban Hydrology. Canadian Water Resources
Journal. 32:193-212.